Self-driving cars: the race to find a business model

By August 10, 2016
Une Tesla modèle S

Wars between automobile manufacturers and startups, alliances between the two worlds, co-operation with universities, a surge in global mobility systems: the self-driving car is on the point of becoming a reality and is the focal point for all kinds of technological and business experiments. We bring you an overview of the situation.

The prestigious National Geographic Magazine ranked self-driving cars among the most promising innovations for the future of mobility in a report it published way back in 1969. The NG journalist opined at the time: “You probably have some concerns about handing over control of your car to a faceless computer. I felt the same until I realised that a computer would also be controlling the other cars. No more reckless drivers swerving across my line or racing towards me head-on. The traffic police and life insurance companies will be delighted to see the unpredictable human element eliminated from our transportation system.”

That article didn’t set much of a precedent at the time. However, a concept that has remained more or less in the realm of science ­fiction seems, forty years on, to be just on the point of becoming a reality. In fact, since 2009 Google Cars have self-driven over 1.5 million miles across the United States. Meanwhile vehicles sold by Tesla are already able to take over part of the driving – though a recent fatal accident has cast something of a shadow over the Tesla on-board system. Los Angeles-based Faraday Future has just received approval  from the California Department of Motor Vehicles to test self-driving cars on the state’s roads. And mysterious white vehicles, laden with cameras and antennae, have recently been spotted on San Francisco’s roads, arousing widespread speculation that they belong to Apple.

The self-driving car as imagined in the 1960s

Watson at the wheel of a self-driving bus?

The self-driving car market thus seems to be hotting up. Since our article in January reporting on the growing number of self-driving car partnerships between start­ups and traditional auto manufacturers, this trend has stepped up another notch. In May alone, Toyota signed a partnership agreement with Uber to provide a new leasing programme for Uber drivers; Volkswagen announced a $300 million investment in Gett – an Israeli-based startup that connects customers with taxi drivers; Apple pumped one billion dollars into Didi Chuxing, the Chinese version of Uber; and Google teamed up with Fiat ­Chrysler to build a fleet of self-driving mini-vans.

The autonomous vehicle developers also seem keen to collaborate with universities, as demonstrated by the partnership between Uber and Carnegie Mellon University and with the University of Arizona. Partnerships are being forged, and new players are also elbowing their way into the sector: San Jose-based startup Next is negotiating with a number of local authorities which are interested in its self-driving transport modules. Arizona-based Local Motors is developing Olli, a self-driving bus that draws on the technology used by IBM supercomputer Watson and can converse with passengers. It is scheduled to be tried out in Washington D.C., starting this summer.

Olli, the self-driving bus

No steering wheel and no pedals

Highly secretive Palo Alto-based start­up Zoox is developing a car that reinvents every aspect of the vehicles with which we have become familiar since the very first car was sold by Karl Benz in 1886. Sources indicate that Zoox creations have neither steering wheel nor pedals – items which Google has already got rid of – boast four drive wheels and will be equally happy to drive either forwards or backwards.

So it seems that the motor car is about to undergo more changes in the next thirty years than it has since it was first created. However, some players reckon that the best way to encourage widespread uptake of self-driving cars among the general public is to sell them a kit designed to turn any ordinary car into a self-driving vehicle. This route is being studied by start­ups Pearl Automation Inc., Geohot and Nauto.

The growing number of partnerships between tech start­ups and traditional auto makers underlines the common interest shared by these very different firms. Their concepts of the self-driving vehicle may differ radically due to their business focus. Young fledgling companies from Silicon Valley tend to regard self-driving cars as computers on wheels whereas their more traditional counterparts see them in a completely different way – as vehicles equipped with computers. But their goals are very similar, i.e. getting self-driving cars on to the roads as soon as possible. Accordingly, in April Google, Uber, Lyft, Ford and Volvo set up a lobbying group called the ‘Self-driving Coalition for Safer Streets’, whose purpose is to help shape US policy and regulation governing autonomous vehicles.

The DARPA (Defense Advanced Research Projects Agency) Grand Challenge, a self-driving car race in the desert, has helped to raise awareness of the technology through the media


Ford versus Google

In trying to achieve their goals, the Internet and ICT specialists have the advantage of being able to take an entirely fresh look at what a ‘car’ is. Given that they are building cars for the first time, they can express themselves more freely without deep-rooted restrictions and ways of working that often make the traditional manufacturers hesitant about innovation.
Moreover, tech start¬ups tend to approach innovation in a way which is at odds with the design and manufacturing culture that arose in the industrial hub based in Detroit, Michigan. While the traditional auto makers are in the habit of introducing innovations incrementally, which leads them to make their way towards self-driving vehicles step by step, young start¬ups have less anxiety and want to get straight to the objective. Ford is currently studying around thirty options on the way towards the semi-autonomous vehicle, including an automatic braking system, whereas Google is building fully self-driving cars from scratch. This is a faster approach, but it is riskier and has not yet stood the test of time.

A Google car.

The illegitimate offspring of a manga character and a computer mouse?

While the tech companies’ lack of auto industry experience can prove a strength in some respects, it definitely has its downside and though start­ups often demonstrate the ability to innovate in bold, radical ways, they are also prone to error. In addition to the recent fatal accident, Tesla has experienced delivery delays on all its models. Nor has everyone been won over by the Google car concept. “It’s as if a character from an old Japanese cartoon had an illegitimate child with a computer mouse,” declared one CNET journalist, after trying out one of the first models.

However, partnerships between traditional manufacturers and newcomers to the sector enable them to combine the reliability that comes from experience with the boldness of youth and benefit from the complementary skillsets of startups and auto makers. Startups have a breadth of experience in computing and software development, while the auto manufacturers know how to build solid, safe vehicles that already enjoy the trust of the general public. Volvo for example promises to accept liability if there is an accident involving one of its future self-driving vehicles, a promise which a startup would find it difficult to keep.

These two elements – state-of-the-art software and safe vehicles – are essential for building the automobile of the future. “Given that the automobile market has high entry barriers, it’s easier for startups to bring their technology to vehicles that already exist,‟ points out Arvind Satyam, a Smart City expert who works for Cisco Systems.



The software battle

These two types of companies are also able to draw on complementary advantages offered by universities: “Unlike companies, which have to generate revenue quickly, universities are not under such pressure and therefore have the freedom to engage in long-term thinking,” underlines Arvind Satyam. “They can moreover reflect on the issues in an overall, objective manner, as they are not obliged to generate any profit. On the other hand, most don’t have an adequate budget to implement their ideas, to build prototypes or run trials. That’s where the bigger companies’ cash flow comes in.” One example is Ford, which is currently testing its self-driving cars on the University of Michigan campus.

However, common interest does not always mean absence of rivalry, and the partnerships being forged by start­ups and auto makers do not necessarily shelter the traditional companies from competition by the tech newcomers. Luca Rigazio, Director of Engineering at the Panasonic Silicon Valley Laboratory, says that “if you look at the value pyramid, the value of the software is far higher than that of the automobile itself. Software has much greater value than hardware‟. Dan Wellers, Global Lead, Digital Futures at SAP, predicts that “software, which will be running everything from navigation to acceleration, plus entertainment – doing far more than human drivers do today – will account for up to 40% of the total value of the vehicle”.

This being so, can auto makers just leave start­ups to handle the software side or should they be developing their own software in order to survive in the long term? Ron DiCarlantonio, CEO of iNAGO, a pioneer in the field of AI systems, favours the second option: “Automobile manufacturers should rapidly create their own smart platform for the car and develop it so as to add significant value for consumers. Toyota is planning to invest $5 billion over the next five years in artificial intelligence. It’s clear that they have chosen to compete with rather than co-operate with the new technology giants. Other firms, such as Fiat­ Chrysler, have gone for the other option, in the belief that they simply have to adopt these services in order to satisfy their customers. This will have positive results in the short term, but negative in the long term. The key thing for firms such as Toyota is that they will need to think more like the new technology firms, to innovate faster and also acquire innovative start­ups that are developing the right technology.‟

Detroit, historically the cradle of the US auto industry and home of traditional automobile manufacturers

Beyond the self-driving car

These new alliances between traditional auto makers and startup software specialists do not all have the one and only aim of building self-driving cars. There are also partnerships involving young on-demand economy companies specialising in car-pooling, whose goal is to move gradually towards an ecosystem of autonomous shared taxis. In addition to substantially reducing the number of road accidents, this new mobility ecosystem would mean fewer – or perhaps even zero – traffic jams, lower greenhouse gas emissions, better usage of urban spaces due to a substantial reduction in the need for parking spaces, increased productivity as people get into the habit of working during their commute, and an easier way for senior citizens and people with disabilities to get around.

Thus a complete rethink of our traditional concept of the car is now under way. The vehicles we know today were designed to be driven by people, mainly for private family use. In a world of shared, self-driving cars, there is no need to have front seats and back seats. All the seats can face each other so that the passengers can talk to each other easily. Larger vehicles could also be designed with tables/desks for people to work, large windows to enable the occupants to admire the view, coffee machines, plus many other facilities. In January 2015 Mercedes­-Benz unveiled a ‘concept car’ based on this kind of innovative approach. Zoox and Next, two pioneering startups mentioned above, are also working in this area. Once people are freed from driving duty, the car’s interior space can be developed as a service plat­form for the passengers. “A new economy will emerge around the user experience inside the car,” predicts Arvind Satyam. “The car will not be thought of simply as just a way of getting from point A to point B, but also as a leisure space. Today, most new cars have WiFi access so that passengers can connect to the Internet and watch videos on YouTube, etc.” We could also imagine vehicles with large screens for viewing content, bringing new opportunities for advertisers.


California, especially San Francisco and Silicon Valley, is home to startups which are designing the cars of the future

Building a coherent ecosystem

This change in our conception of the automobile and how we use it can already be seen in the success of ride-sharing companies such as Uber, Lyft and Blablacar. Argues Arvind Satyam: “Consumers don’t necessarily want to own a car, they want to benefit from a service which gets them around.‟  This trend stems from the surge in the subscription economy and the on-demand economy, the thinking being that the most important thing is not to own an asset but to be able to use the services it provides. Thus many people find it more convenient to use a mobile app to order a car to go to a given destination than to pay all the costs associated with owning a private car, which will then most probably spend most of its time in the garage.

This trend looks likely to grow. In terms of the business proposition this would mean a shift of emphasis away from building and selling cars towards taking customers where they want to go. General Motors has grasped this idea, and has teamed up with Lyft and Cruise Automation. General Motors’ vehicles, Lyft ride-sharing service and Cruise Automation’s technology are laying the foundations of an ecosystem that provides an efficient, easily available ‘mobility’ service for users. Meanwhile Ford launched a ‘Smart Mobility’ subsidiary of its own in March this year.

In this context, private sector specialists ought to be working closely with local authorities to develop fully joined-up transportation systems. “Today I know that I can take an Uber or the Caltrain to Palo Alto, for instance, but there still isn’t a solution that uses all the possible combinations and takes account of my personal preferences so as to offer me an optimised, personalised itinerary,” points out Arvind Satyam, arguing: “That’s where the local authorities come in. They have access to huge amounts of data and can view things in a holistic way. If they were to build a streamlined platform bringing together all the data in a coherent manner, they could then instruct private companies about which problems needed solving as a priority.‟

Moving towards an ecosystem that links up all the various means of transportation?

The Phantom Auto, forerunner to the autonomous car

In the inter-war years, a number of driverless vehicles were already plying the roads of the United States, sometimes creating amazement among the crowds that gathered to watch them pass by. These vehicles were not controlled by computer – computers did not exist at that time – nor were they truly autonomous: a flesh-and-blood operator sitting in a second vehicle used to pilot the car remotely, but that in no way diminished the excitement among the spectators. A column written by a gushing journalist at the Free Lance Star in Fredericksburg, Virginia gives a flavour of this excitement: “No one touching it, no wires or strings attached to it, weaving in and out of traffic, climbing hills, turning corners, stopping for traffic lights, just as though there were an invisible driver at the wheel!” The ‘Times Recorder’ in Zanesville, Ohio also reported on a vehicle controlled by a driver who was perched in a low-flying aeroplane, calling it “one of the most amazing products of modern science.”

However, it turned out not to be one of the safest. In 1932 an accident in which a driverless car mowed down a number of people in Hanover, Pennsylvania caused a distinct cooling in the mass enthusiasm for remotely-piloted vehicles. Then science fiction took over. The German writer Werner Illing wrote in his 1930 novel Utopolis of vehicles “behaving as if they had learned the Highway Code by heart.” Five years later, in his novel ‘The Living Machine’, American writer David H. Keller listed the advantages of self-driving cars: ‟Old people began to cross the continent in their own cars. Young people found the driverless car admirable for petting. The blind for the first time were safe. Parents found they could more safely send their children to school in the new car than in the old cars with a chauffeur”. In ’Sally’, a science fiction short story, Isaac Asimov portrays a world where only autonomous vehicles are authorised to use the roads. The films Total Recall, Demolition Man, Minority Report and I, Robot have all also brought driverless vehicles to the big screen. Now however, as so often before, reality is beginning to overtake fiction.


Legal mentions © L’Atelier BNP Paribas